One of the most widespread disorders in childhood is called autism spectrum disorder (ASD), which affects the brain's function. Previously, many efforts have been made to develop an intelligent system to detect disease using brain activity. However, accurate diagnosis of ASD remains a challenging issue among scientists. The purpose of this study was to diagnose ASD at an early age using a low computationally algorithm based on electroencephalography (EEG) signal. In this study, we classified two groups of normal and autistic children using brain signals at resting-state. Two brain channels (C3 and C4) of 61 children including 27 normal children and 34 autistic children in the age range of 4 to 8 years were studied. For the first time, we characterized the EEGs using innovative polar-based lagged state-space indices. The classification was performed using the support vector machine (SVM). The results demonstrated the highest average accuracy of 81.96% using the indices of two EEG channels. Using single-channel EEG measures, the maximum average classification rate of 78.68% was achieved using C4. To sum up, the results revealed that despite the limited number of brain channels and computational simplicity, the proposed algorithm was able to distinguish the two groups of normal and autistic children with satisfactory accuracy.

儿童期最普遍的疾病之一称为自闭症谱系障碍（ASD），它会影响大脑的功能。以前，已经做出了许多努力来开发一种智能系统，该系统可以利用大脑活动来检测疾病。但是，准确诊断ASD仍然是科学家们面临的挑战性问题。这项研究的目的是使用基于脑电图（EEG）信号的低计算算法来诊断ASD。在这项研究中，我们使用静止状态下的脑信号将正常儿童和自闭症儿童分为两组。研究了61名儿童的两个大脑通道（C3和C4），包括27名正常儿童和34名4至8岁的自闭症儿童。我们首次使用创新的基于极性的滞后状态空间指数来表征脑电图。使用支持向量机（SVM）进行分类。结果表明，使用两个EEG通道的指数，最高平均准确度为81.96％。使用单通道脑电图测量，使用C4可以达到78.68％的最大平均分类率。综上所述，结果表明，尽管大脑通道数量有限且计算简单，但该算法仍能够以令人满意的准确度区分正常儿童和自闭症儿童。